LabelScope: Context-Aware Discovery of Label Outliers in Radio Astronomy Signals

LabelScope: Context-Aware Discovery of Label Outliers in Radio Astronomy Signals. Hossain, R., Mason, B., Loomis, R., Phillips, J. M., & Rezig, E. K. Proceedings of the VLDB Endowment, 19(12):4686–4689, August, 2026.

Paper doi abstract bibtex

The National Radio Astronomy Observatory (NRAO) operates one of the largest radio telescopes in the world. As part of its data processing pipeline, it produces thousands of bandpass calibration signals per observation cycle, each capturing how an antenna’s frequency response varies across a spectral window—essentially a sequence of amplitude measurements over frequency channels. To ensure data quality, automated detectors assign anomaly scores to these signals, and human experts review flagged cases to assign labels. In practice, scores and labels can disagree, and expert review cannot scale to every mismatch. We present LabelScope, an interactive system developed in collaboration with NRAO to address this problem. LabelScope groups signals using contextual metadata and feature augmentation, identifies minority-label instances within locally coherent groups, and generates predicate-based explanations for why they are interesting disagreement cases. Through its interface, users explore candidate label outliers and determine whether score–label disagreements reflect labeling errors, meaningful anomalies, or limitations of the scoring method. We demonstrate LabelScope on real astronomy signals, enabling NRAO scientists to efficiently direct limited reviewer attention toward the most informative disagreements. Beyond astronomy, this demonstration exposes VLDB 2026 attendees to a class of real-world data quality problems, i.e., contextual disagreements between automated anomaly scoring and human labeling, that arises broadly in scientific data pipelines but remains underexplored in the data management literature.

@article{hossain_labelscope_2026,
	title = {{LabelScope}: {Context}-{Aware} {Discovery} of {Label} {Outliers} in {Radio} {Astronomy} {Signals}},
	volume = {19},
	issn = {2150-8097},
	shorttitle = {{LabelScope}},
	url = {https://dl.acm.org/doi/10.14778/3827998.3828097},
	doi = {10.14778/3827998.3828097},
	abstract = {The National Radio Astronomy Observatory (NRAO) operates one of the largest radio telescopes in the world. As part of its data processing pipeline, it produces thousands of bandpass calibration signals per observation cycle, each capturing how an antenna’s frequency response varies across a spectral window—essentially a sequence of amplitude measurements over frequency channels. To ensure data quality, automated detectors assign anomaly scores to these signals, and human experts review flagged cases to assign labels. In practice, scores and labels can disagree, and expert review cannot scale to every mismatch. We present LabelScope, an interactive system developed in collaboration with NRAO to address this problem. LabelScope groups signals using contextual metadata and feature augmentation, identifies minority-label instances within locally coherent groups, and generates predicate-based explanations for why they are interesting disagreement cases. Through its interface, users explore candidate label outliers and determine whether score–label disagreements reflect labeling errors, meaningful anomalies, or limitations of the scoring method. We demonstrate LabelScope on real astronomy signals, enabling NRAO scientists to efficiently direct limited reviewer attention toward the most informative disagreements. Beyond astronomy, this demonstration exposes VLDB 2026 attendees to a class of real-world data quality problems, i.e., contextual disagreements between automated anomaly scoring and human labeling, that arises broadly in scientific data pipelines but remains underexplored in the data management literature.},
	language = {en},
	number = {12},
	urldate = {2026-09-16},
	journal = {Proceedings of the VLDB Endowment},
	author = {Hossain, Rabeya and Mason, Brian and Loomis, Ryan and Phillips, Jeff M. and Rezig, El Kindi},
	month = aug,
	year = {2026},
	pages = {4686--4689},
}

Downloads: 0

{"_id":"qnaHwT4gFQbjCHAvg","bibbaseid":"hossain-mason-loomis-phillips-rezig-labelscopecontextawarediscoveryoflabeloutliersinradioastronomysignals-2026","author_short":["Hossain, R.","Mason, B.","Loomis, R.","Phillips, J. M.","Rezig, E. K."],"bibdata":{"bibtype":"article","type":"article","title":"LabelScope: Context-Aware Discovery of Label Outliers in Radio Astronomy Signals","volume":"19","issn":"2150-8097","shorttitle":"LabelScope","url":"https://dl.acm.org/doi/10.14778/3827998.3828097","doi":"10.14778/3827998.3828097","abstract":"The National Radio Astronomy Observatory (NRAO) operates one of the largest radio telescopes in the world. As part of its data processing pipeline, it produces thousands of bandpass calibration signals per observation cycle, each capturing how an antenna’s frequency response varies across a spectral window—essentially a sequence of amplitude measurements over frequency channels. To ensure data quality, automated detectors assign anomaly scores to these signals, and human experts review flagged cases to assign labels. In practice, scores and labels can disagree, and expert review cannot scale to every mismatch. We present LabelScope, an interactive system developed in collaboration with NRAO to address this problem. LabelScope groups signals using contextual metadata and feature augmentation, identifies minority-label instances within locally coherent groups, and generates predicate-based explanations for why they are interesting disagreement cases. Through its interface, users explore candidate label outliers and determine whether score–label disagreements reflect labeling errors, meaningful anomalies, or limitations of the scoring method. We demonstrate LabelScope on real astronomy signals, enabling NRAO scientists to efficiently direct limited reviewer attention toward the most informative disagreements. Beyond astronomy, this demonstration exposes VLDB 2026 attendees to a class of real-world data quality problems, i.e., contextual disagreements between automated anomaly scoring and human labeling, that arises broadly in scientific data pipelines but remains underexplored in the data management literature.","language":"en","number":"12","urldate":"2026-09-16","journal":"Proceedings of the VLDB Endowment","author":[{"propositions":[],"lastnames":["Hossain"],"firstnames":["Rabeya"],"suffixes":[]},{"propositions":[],"lastnames":["Mason"],"firstnames":["Brian"],"suffixes":[]},{"propositions":[],"lastnames":["Loomis"],"firstnames":["Ryan"],"suffixes":[]},{"propositions":[],"lastnames":["Phillips"],"firstnames":["Jeff","M."],"suffixes":[]},{"propositions":[],"lastnames":["Rezig"],"firstnames":["El","Kindi"],"suffixes":[]}],"month":"August","year":"2026","pages":"4686–4689","bibtex":"@article{hossain_labelscope_2026,\n\ttitle = {{LabelScope}: {Context}-{Aware} {Discovery} of {Label} {Outliers} in {Radio} {Astronomy} {Signals}},\n\tvolume = {19},\n\tissn = {2150-8097},\n\tshorttitle = {{LabelScope}},\n\turl = {https://dl.acm.org/doi/10.14778/3827998.3828097},\n\tdoi = {10.14778/3827998.3828097},\n\tabstract = {The National Radio Astronomy Observatory (NRAO) operates one of the largest radio telescopes in the world. As part of its data processing pipeline, it produces thousands of bandpass calibration signals per observation cycle, each capturing how an antenna’s frequency response varies across a spectral window—essentially a sequence of amplitude measurements over frequency channels. To ensure data quality, automated detectors assign anomaly scores to these signals, and human experts review flagged cases to assign labels. In practice, scores and labels can disagree, and expert review cannot scale to every mismatch. We present LabelScope, an interactive system developed in collaboration with NRAO to address this problem. LabelScope groups signals using contextual metadata and feature augmentation, identifies minority-label instances within locally coherent groups, and generates predicate-based explanations for why they are interesting disagreement cases. Through its interface, users explore candidate label outliers and determine whether score–label disagreements reflect labeling errors, meaningful anomalies, or limitations of the scoring method. We demonstrate LabelScope on real astronomy signals, enabling NRAO scientists to efficiently direct limited reviewer attention toward the most informative disagreements. Beyond astronomy, this demonstration exposes VLDB 2026 attendees to a class of real-world data quality problems, i.e., contextual disagreements between automated anomaly scoring and human labeling, that arises broadly in scientific data pipelines but remains underexplored in the data management literature.},\n\tlanguage = {en},\n\tnumber = {12},\n\turldate = {2026-09-16},\n\tjournal = {Proceedings of the VLDB Endowment},\n\tauthor = {Hossain, Rabeya and Mason, Brian and Loomis, Ryan and Phillips, Jeff M. and Rezig, El Kindi},\n\tmonth = aug,\n\tyear = {2026},\n\tpages = {4686--4689},\n}\n\n\n\n\n\n\n\n\n\n\n\n","author_short":["Hossain, R.","Mason, B.","Loomis, R.","Phillips, J. M.","Rezig, E. K."],"key":"hossain_labelscope_2026","id":"hossain_labelscope_2026","bibbaseid":"hossain-mason-loomis-phillips-rezig-labelscopecontextawarediscoveryoflabeloutliersinradioastronomysignals-2026","role":"author","urls":{"Paper":"https://dl.acm.org/doi/10.14778/3827998.3828097"},"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://bibbase.org/zotero-group/pratikmhatre/5933976","dataSources":["yJr5AAtJ5Sz3Q4WT4"],"keywords":[],"search_terms":["labelscope","context","aware","discovery","label","outliers","radio","astronomy","signals","hossain","mason","loomis","phillips","rezig"],"title":"LabelScope: Context-Aware Discovery of Label Outliers in Radio Astronomy Signals","year":2026}